

My Journey
Inspired by Dr. Asim Karim, a blind professor in assistive technology, and Conor Russomanno, CEO of OpenBCI, I ventured into BCI research to revolutionize not just speech recovery but broader applications in brain-computer interaction. With a passion for AI and robotics, I founded Excelleve, a BCI-driven assistive technology company, to push the boundaries of innovation and create life-changing solutions.
Projects
๐ง NeuroAI & Brain Signal Analysis
๐น Thought-to-Speech BCI
Objective: Restore speech for individuals with motor impairments.
Technologies Used: EEG, Deep Learning, Signal Processing.
Impact: Developing non-invasive solutions for speech restoration with broader applications in Brain-Computer Interfaces (BCI).
๐น Harmful Brain Activity Detection
Applied deep learning to EEG data to classify neurological disorders such as LPD and seizures, enabling early diagnosis and detection.
๐ฉบ Medical AI & Imaging
๐น Medical Image Segmentation
Developed AI-driven models to segment and classify medical anomalies in MRI, CT scans, and X-ray imaging.
Impact: Enhanced accuracy and speed in medical diagnostics through automated region-based analysis.
๐น Disease Classification from Medical Imaging
Built AI models to analyze and classify radiological scans, improving diagnostic accuracy.
Impact: Demonstrated potential to enhance diagnostic accuracy and support clinical decision-making in research settings.
๐น AI-Powered Disease Prediction
Created predictive AI models using structured patient data (e.g., lab results, history) for early detection of critical diseases.
Impact: Supported preventative care and medical decision-making by identifying high-risk patients.
๐ฟ AI for Ecology & Conservation
๐น WildNet: AI for Bird Species Recognition
Objective: Classify bird species from environmental audio recordings to support large-scale biodiversity monitoring and wildlife conservation.
Technologies Used: Custom deep learning architecture combining Convolutional Neural Networks (CNNs), channel attention mechanisms, and transformer encoders for multi-label audio classification.
Impact: Advanced ecological research by enabling accurate species identification in diverse acoustic environments, contributing to conservation and biodiversity tracking efforts.
๐ค Advanced Robotics & IoT
๐น Robotics & Automation
Designed and implemented high-precision robotic systems; awarded in national and international robotics competitions.
๐น Embedded Systems & Control
Engineered robotics solutions using PID control, sensor integration, and ArduinoโVEX systems for intelligent automation.
๐น IoT & Smart Systems
Developed IoT-based applications using microcontrollers and embedded systems for real-world automation tasks.
๐น Circuit Design & PCB Development
Designed custom PCBs for robotic and IoT systems, improving performance, reliability, and integration.
๐น Robotics Mentorship & R&D Leadership
Led student teams in robotics competitions, mentoring in advanced robotics, embedded systems, and mechatronics.
๐ฅ Generative AI & Large Language Models (LLMs)
๐น Fine-Tuning LLMs for Specialized Applications
Optimized transformer-based architectures for domain-specific tasks such as Olympiad-level math solving and AI-generated content classification.
๐น Automated RLHF in LLM Training
Developed a system to automate Reinforcement Learning from Human Feedback (RLHF), reducing reliance on manual tuning and accelerating LLM optimization.
๐น AI-Driven Workflow Automation
Built AI solutions to automate repetitive tasks, enhance productivity, and streamline operations in real-world environments.
Get In Touch
Muhammad Huzyafa Khokhar
muhammadhuzyafakhokhar@gmail.com

